Cochlear implant hearing assistance devices have been known in many years in a variety of configurations, but typically comprising
a) a number of electrodes implantable in different locations of the cochlea allowing a stimulation of different frequencies of the audible range,
b) an external part for picking up and processing sound from the environment, and for determining sequences of pulses for stimulation of the electrodes in dependence on the current input sound,
c) a (typically wireless, e.g. inductive) communication link for simultaneously transmitting information about the stimulation sequences and for transferring energy to
d) an implanted part allowing the stimulation to be generated and applied to the relevant of said electrodes.
Such systems are e.g. described in U.S. Pat. No. 4,207,441 and in U.S. Pat. No. 4,532,930.
To adapt a cochlear implant type hearing assistance device to a user's particular needs, information about the users' hearing threshold (T) and comfort (C) levels of electrical stimulation intensity as a function of frequency is needed. During a fitting session, electrically evoked compound potentials (eCAPs) in response to various levels of electrical stimulation signals can e.g. be measured for different electrodes stimulating different parts of the auditory nerve. In a subsequent processing procedure, the thus recorded signals can be used to extract the actual electrical nerve responses from the user's nerve cells and by mapping corresponding values of the amplitude of the nerve response signal AeCAP versus the intensity (energy) of the stimulation signal Is, a ‘hearing threshold’ (T) stimulation intensity (IT) for each electrode can be determined.
Nerve response measurements (e.g. eCAP-measurements) and their subsequent processing can be performed in a number of different ways, e.g. (in a relatively ‘simple’ way) by applying electric stimuli to an implanted electrode and using external ABR electrodes (surface pick-up electrodes applied to the skin) to sense the response of the human auditory system. The first human recordings of eCAPs were published by [Brown et al.; 1990] using a method published by [de Sauvage et al.; 1983].
Electrically evoked compound action potentials (eCAPs) are routinely used in clinical audiology to set the threshold level of a patient's cochlear implant processor. The separation of nerve responses from artifacts created by the stimulation signals and estimation of hearing thresholds from evoked potentials have in the past been carried out in a variety of different ways, e.g. often involving non-automatic procedure steps, e.g. the judgment of experts. Such non-automatic procedures can be time consuming and prone to errors. A review of efforts to identify and remove noise in eCAP-measurements has been published by [Undurraga et al.; 2012].
In general, measurement of eCAPs are used                During an operation, where a cochlear implant part of a hearing assistance device is inserted, to test whether the stimulation of the cochlear implant is functioning well AND whether the nerve is responding. These facts must be verified before the operation is ended, hence requiring a reliable and fast measurement and verification procedure to minimize operation time.        During fitting to determine hearing thresholds. Also here a fast and reliable method is preferable (although the time pressure is slightly less severe).        
Thus there is a need for a reliable and fast method of identifying and processing electrically evoked (hearing) nerve responses.